stringex 1.0.0

data/test/unidecoder_test.rb

@@ -0,0 +1,90 @@
+require "test/unit"
+
+$: << File.join(File.dirname(__FILE__), '../lib')
+require File.join(File.dirname(__FILE__), "../init")
+
+class UnidecoderTest < Test::Unit::TestCase
+  # Silly phrases courtesy of Frank da Cruz
+  # http://www.columbia.edu/kermit/utf8.html
+  DONT_CONVERT = [
+    "Vitrum edere possum; mihi non nocet.", # Latin
+    "Je puis mangier del voirre. Ne me nuit.", # Old French
+    "Kristala jan dezaket, ez dit minik ematen.", # Basque
+    "Kaya kong kumain nang bubog at hindi ako masaktan.", # Tagalog
+    "Ich kann Glas essen, ohne mir weh zu tun.", # German
+    "I can eat glass and it doesn't hurt me.", # English
+  ]
+  CONVERT_PAIRS = {
+    "Je peux manger du verre, ça ne me fait pas de mal." => # French
+      "Je peux manger du verre, ca ne me fait pas de mal.",
+    "Pot să mănânc sticlă și ea nu mă rănește." => # Romanian
+      "Pot sa mananc sticla si ea nu ma raneste.",
+    "Ég get etið gler án þess að meiða mig." => # Icelandic
+      "Eg get etid gler an thess ad meida mig.",
+    "Unë mund të ha qelq dhe nuk më gjen gjë." => # Albanian
+      "Une mund te ha qelq dhe nuk me gjen gje.",
+    "Mogę jeść szkło i mi nie szkodzi." => # Polish
+      "Moge jesc szklo i mi nie szkodzi.",
+    "Я могу есть стекло, оно мне не вредит." => # Russian
+      "Ia moghu iest' stieklo, ono mnie nie vriedit.",
+    "Мога да ям стъкло, то не ми вреди." => # Bulgarian
+      "Mogha da iam stklo, to nie mi vriedi.",
+    "ᛁᚳ᛫ᛗᚨᚷ᛫ᚷᛚᚨᛋ᛫ᛖᚩᛏᚪᚾ᛫ᚩᚾᛞ᛫ᚻᛁᛏ᛫ᚾᛖ᛫ᚻᛖᚪᚱᛗᛁᚪᚧ᛫ᛗᛖ᛬" => # Anglo-Saxon
+      "ic.mag.glas.eotacn.ond.hit.ne.heacrmiacth.me:",
+    "ὕαλον ϕαγεῖν δύναμαι· τοῦτο οὔ με βλάπτει" => # Classical Greek
+      "ualon phagein dunamai; touto ou me blaptei",
+    "मैं काँच खा सकता हूँ और मुझे उससे कोई चोट नहीं पहुंचती" => # Hindi
+      "maiN kaaNc khaa sktaa huuN aur mujhe usse koii cott nhiiN phuNctii",
+    "من می توانم بدونِ احساس درد شيشه بخورم" => # Persian
+      "mn my twnm bdwni Hss drd shyshh bkhwrm",
+    "أنا قادر على أكل الزجاج و هذا لا يؤلمن" => # Arabic
+      "'n qdr `l~ 'kl lzjj w hdh l yw'lmn",
+    "אני יכול לאכול זכוכית וזה לא מזיק לי" => # Hebrew
+      "ny ykvl lkvl zkvkyt vzh l mzyq ly",
+    "ฉันกินกระจกได้ แต่มันไม่ทำให้ฉันเจ็บ" => # Thai
+      "chankinkracchkaid aetmanaimthamaihchanecchb",
+    "我能吞下玻璃而不伤身体。" => # Chinese
+      "Wo Neng Tun Xia Bo Li Er Bu Shang Shen Ti . ",
+    "私はガラスを食べられます。それは私を傷つけません。" => # Japanese
+      "Si hagarasuwoShi beraremasu. sorehaSi woShang tukemasen. "
+  }
+  
+  def test_unidecoder_decode
+    DONT_CONVERT.each do |ascii|
+      assert_equal ascii, LuckySneaks::Unidecoder.decode(ascii)
+    end
+    CONVERT_PAIRS.each do |unicode, ascii|
+      assert_equal ascii, LuckySneaks::Unidecoder.decode(unicode)
+    end
+  end
+  
+  def test_to_ascii
+    DONT_CONVERT.each do |ascii|
+      assert_equal ascii, ascii.to_ascii
+    end
+    CONVERT_PAIRS.each do |unicode, ascii|
+      assert_equal ascii, unicode.to_ascii
+    end
+  end
+  
+  def test_unidecoder_encode
+    {
+      # Strings
+      "0041" => "A",
+      "00e6" => "æ",
+      "042f" => "Я"
+    }.each do |codepoint, unicode|
+      assert_equal unicode, LuckySneaks::Unidecoder.encode(codepoint)
+    end
+  end
+  
+  def test_unidecoder_in_yaml_file
+    {
+      "A" => "x00.yml (line 67)",
+      "π" => "x03.yml (line 194)",
+      "Я" => "x04.yml (line 49)"
+    }.each do |character, output|
+      assert_equal output, LuckySneaks::Unidecoder.in_yaml_file(character)
+    end
+  end
+end

metadata

@@ -0,0 +1,250 @@
+--- !ruby/object:Gem::Specification 
+name: stringex
+version: !ruby/object:Gem::Version 
+  version: 1.0.0
+platform: ruby
+authors: 
+- Russell Norris
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2009-07-20 00:00:00 -04:00
+default_executable: 
+dependencies: []
+
+description: "Some [hopefully] useful extensions to Ruby\xE2\x80\x99s String class. Stringex is made up of three libraries: ActsAsUrl [permalink solution with better character translation], Unidecoder [Unicode to Ascii transliteration], and StringExtensions [miscellaneous helper methods for the String class]."
+email: rsl@luckysneaks.com
+executables: []
+
+extensions: []
+
+extra_rdoc_files: 
+- MIT-LICENSE
+- README.rdoc
+files: 
+- MIT-LICENSE
+- README.rdoc
+- Rakefile
+- init.rb
+- lib/lucky_sneaks/acts_as_url.rb
+- lib/lucky_sneaks/string_extensions.rb
+- lib/lucky_sneaks/unidecoder.rb
+- lib/lucky_sneaks/unidecoder_data/x00.yml
+- lib/lucky_sneaks/unidecoder_data/x01.yml
+- lib/lucky_sneaks/unidecoder_data/x02.yml
+- lib/lucky_sneaks/unidecoder_data/x03.yml
+- lib/lucky_sneaks/unidecoder_data/x04.yml
+- lib/lucky_sneaks/unidecoder_data/x05.yml
+- lib/lucky_sneaks/unidecoder_data/x06.yml
+- lib/lucky_sneaks/unidecoder_data/x07.yml
+- lib/lucky_sneaks/unidecoder_data/x09.yml
+- lib/lucky_sneaks/unidecoder_data/x0a.yml
+- lib/lucky_sneaks/unidecoder_data/x0b.yml
+- lib/lucky_sneaks/unidecoder_data/x0c.yml
+- lib/lucky_sneaks/unidecoder_data/x0d.yml
+- lib/lucky_sneaks/unidecoder_data/x0e.yml
+- lib/lucky_sneaks/unidecoder_data/x0f.yml
+- lib/lucky_sneaks/unidecoder_data/x10.yml
+- lib/lucky_sneaks/unidecoder_data/x11.yml
+- lib/lucky_sneaks/unidecoder_data/x12.yml
+- lib/lucky_sneaks/unidecoder_data/x13.yml
+- lib/lucky_sneaks/unidecoder_data/x14.yml
+- lib/lucky_sneaks/unidecoder_data/x15.yml
+- lib/lucky_sneaks/unidecoder_data/x16.yml
+- lib/lucky_sneaks/unidecoder_data/x17.yml
+- lib/lucky_sneaks/unidecoder_data/x18.yml
+- lib/lucky_sneaks/unidecoder_data/x1e.yml
+- lib/lucky_sneaks/unidecoder_data/x1f.yml
+- lib/lucky_sneaks/unidecoder_data/x20.yml
+- lib/lucky_sneaks/unidecoder_data/x21.yml
+- lib/lucky_sneaks/unidecoder_data/x22.yml
+- lib/lucky_sneaks/unidecoder_data/x23.yml
+- lib/lucky_sneaks/unidecoder_data/x24.yml
+- lib/lucky_sneaks/unidecoder_data/x25.yml
+- lib/lucky_sneaks/unidecoder_data/x26.yml
+- lib/lucky_sneaks/unidecoder_data/x27.yml
+- lib/lucky_sneaks/unidecoder_data/x28.yml
+- lib/lucky_sneaks/unidecoder_data/x2e.yml
+- lib/lucky_sneaks/unidecoder_data/x2f.yml
+- lib/lucky_sneaks/unidecoder_data/x30.yml
+- lib/lucky_sneaks/unidecoder_data/x31.yml
+- lib/lucky_sneaks/unidecoder_data/x32.yml
+- lib/lucky_sneaks/unidecoder_data/x33.yml
+- lib/lucky_sneaks/unidecoder_data/x4d.yml
+- lib/lucky_sneaks/unidecoder_data/x4e.yml
+- lib/lucky_sneaks/unidecoder_data/x4f.yml
+- lib/lucky_sneaks/unidecoder_data/x50.yml
+- lib/lucky_sneaks/unidecoder_data/x51.yml
+- lib/lucky_sneaks/unidecoder_data/x52.yml
+- lib/lucky_sneaks/unidecoder_data/x53.yml
+- lib/lucky_sneaks/unidecoder_data/x54.yml
+- lib/lucky_sneaks/unidecoder_data/x55.yml
+- lib/lucky_sneaks/unidecoder_data/x56.yml
+- lib/lucky_sneaks/unidecoder_data/x57.yml
+- lib/lucky_sneaks/unidecoder_data/x58.yml
+- lib/lucky_sneaks/unidecoder_data/x59.yml
+- lib/lucky_sneaks/unidecoder_data/x5a.yml
+- lib/lucky_sneaks/unidecoder_data/x5b.yml
+- lib/lucky_sneaks/unidecoder_data/x5c.yml
+- lib/lucky_sneaks/unidecoder_data/x5d.yml
+- lib/lucky_sneaks/unidecoder_data/x5e.yml
+- lib/lucky_sneaks/unidecoder_data/x5f.yml
+- lib/lucky_sneaks/unidecoder_data/x60.yml
+- lib/lucky_sneaks/unidecoder_data/x61.yml
+- lib/lucky_sneaks/unidecoder_data/x62.yml
+- lib/lucky_sneaks/unidecoder_data/x63.yml
+- lib/lucky_sneaks/unidecoder_data/x64.yml
+- lib/lucky_sneaks/unidecoder_data/x65.yml
+- lib/lucky_sneaks/unidecoder_data/x66.yml
+- lib/lucky_sneaks/unidecoder_data/x67.yml
+- lib/lucky_sneaks/unidecoder_data/x68.yml
+- lib/lucky_sneaks/unidecoder_data/x69.yml
+- lib/lucky_sneaks/unidecoder_data/x6a.yml
+- lib/lucky_sneaks/unidecoder_data/x6b.yml
+- lib/lucky_sneaks/unidecoder_data/x6c.yml
+- lib/lucky_sneaks/unidecoder_data/x6d.yml
+- lib/lucky_sneaks/unidecoder_data/x6e.yml
+- lib/lucky_sneaks/unidecoder_data/x6f.yml
+- lib/lucky_sneaks/unidecoder_data/x70.yml
+- lib/lucky_sneaks/unidecoder_data/x71.yml
+- lib/lucky_sneaks/unidecoder_data/x72.yml
+- lib/lucky_sneaks/unidecoder_data/x73.yml
+- lib/lucky_sneaks/unidecoder_data/x74.yml
+- lib/lucky_sneaks/unidecoder_data/x75.yml
+- lib/lucky_sneaks/unidecoder_data/x76.yml
+- lib/lucky_sneaks/unidecoder_data/x77.yml
+- lib/lucky_sneaks/unidecoder_data/x78.yml
+- lib/lucky_sneaks/unidecoder_data/x79.yml
+- lib/lucky_sneaks/unidecoder_data/x7a.yml
+- lib/lucky_sneaks/unidecoder_data/x7b.yml
+- lib/lucky_sneaks/unidecoder_data/x7c.yml
+- lib/lucky_sneaks/unidecoder_data/x7d.yml
+- lib/lucky_sneaks/unidecoder_data/x7e.yml
+- lib/lucky_sneaks/unidecoder_data/x7f.yml
+- lib/lucky_sneaks/unidecoder_data/x80.yml
+- lib/lucky_sneaks/unidecoder_data/x81.yml
+- lib/lucky_sneaks/unidecoder_data/x82.yml
+- lib/lucky_sneaks/unidecoder_data/x83.yml
+- lib/lucky_sneaks/unidecoder_data/x84.yml
+- lib/lucky_sneaks/unidecoder_data/x85.yml
+- lib/lucky_sneaks/unidecoder_data/x86.yml
+- lib/lucky_sneaks/unidecoder_data/x87.yml
+- lib/lucky_sneaks/unidecoder_data/x88.yml
+- lib/lucky_sneaks/unidecoder_data/x89.yml
+- lib/lucky_sneaks/unidecoder_data/x8a.yml
+- lib/lucky_sneaks/unidecoder_data/x8b.yml
+- lib/lucky_sneaks/unidecoder_data/x8c.yml
+- lib/lucky_sneaks/unidecoder_data/x8d.yml
+- lib/lucky_sneaks/unidecoder_data/x8e.yml
+- lib/lucky_sneaks/unidecoder_data/x8f.yml
+- lib/lucky_sneaks/unidecoder_data/x90.yml
+- lib/lucky_sneaks/unidecoder_data/x91.yml
+- lib/lucky_sneaks/unidecoder_data/x92.yml
+- lib/lucky_sneaks/unidecoder_data/x93.yml
+- lib/lucky_sneaks/unidecoder_data/x94.yml
+- lib/lucky_sneaks/unidecoder_data/x95.yml
+- lib/lucky_sneaks/unidecoder_data/x96.yml
+- lib/lucky_sneaks/unidecoder_data/x97.yml
+- lib/lucky_sneaks/unidecoder_data/x98.yml
+- lib/lucky_sneaks/unidecoder_data/x99.yml
+- lib/lucky_sneaks/unidecoder_data/x9a.yml
+- lib/lucky_sneaks/unidecoder_data/x9b.yml
+- lib/lucky_sneaks/unidecoder_data/x9c.yml
+- lib/lucky_sneaks/unidecoder_data/x9d.yml
+- lib/lucky_sneaks/unidecoder_data/x9e.yml
+- lib/lucky_sneaks/unidecoder_data/x9f.yml
+- lib/lucky_sneaks/unidecoder_data/xa0.yml
+- lib/lucky_sneaks/unidecoder_data/xa1.yml
+- lib/lucky_sneaks/unidecoder_data/xa2.yml
+- lib/lucky_sneaks/unidecoder_data/xa3.yml
+- lib/lucky_sneaks/unidecoder_data/xa4.yml
+- lib/lucky_sneaks/unidecoder_data/xac.yml
+- lib/lucky_sneaks/unidecoder_data/xad.yml
+- lib/lucky_sneaks/unidecoder_data/xae.yml
+- lib/lucky_sneaks/unidecoder_data/xaf.yml
+- lib/lucky_sneaks/unidecoder_data/xb0.yml
+- lib/lucky_sneaks/unidecoder_data/xb1.yml
+- lib/lucky_sneaks/unidecoder_data/xb2.yml
+- lib/lucky_sneaks/unidecoder_data/xb3.yml
+- lib/lucky_sneaks/unidecoder_data/xb4.yml
+- lib/lucky_sneaks/unidecoder_data/xb5.yml
+- lib/lucky_sneaks/unidecoder_data/xb6.yml
+- lib/lucky_sneaks/unidecoder_data/xb7.yml
+- lib/lucky_sneaks/unidecoder_data/xb8.yml
+- lib/lucky_sneaks/unidecoder_data/xb9.yml
+- lib/lucky_sneaks/unidecoder_data/xba.yml
+- lib/lucky_sneaks/unidecoder_data/xbb.yml
+- lib/lucky_sneaks/unidecoder_data/xbc.yml
+- lib/lucky_sneaks/unidecoder_data/xbd.yml
+- lib/lucky_sneaks/unidecoder_data/xbe.yml
+- lib/lucky_sneaks/unidecoder_data/xbf.yml
+- lib/lucky_sneaks/unidecoder_data/xc0.yml
+- lib/lucky_sneaks/unidecoder_data/xc1.yml
+- lib/lucky_sneaks/unidecoder_data/xc2.yml
+- lib/lucky_sneaks/unidecoder_data/xc3.yml
+- lib/lucky_sneaks/unidecoder_data/xc4.yml
+- lib/lucky_sneaks/unidecoder_data/xc5.yml
+- lib/lucky_sneaks/unidecoder_data/xc6.yml
+- lib/lucky_sneaks/unidecoder_data/xc7.yml
+- lib/lucky_sneaks/unidecoder_data/xc8.yml
+- lib/lucky_sneaks/unidecoder_data/xc9.yml
+- lib/lucky_sneaks/unidecoder_data/xca.yml
+- lib/lucky_sneaks/unidecoder_data/xcb.yml
+- lib/lucky_sneaks/unidecoder_data/xcc.yml
+- lib/lucky_sneaks/unidecoder_data/xcd.yml
+- lib/lucky_sneaks/unidecoder_data/xce.yml
+- lib/lucky_sneaks/unidecoder_data/xcf.yml
+- lib/lucky_sneaks/unidecoder_data/xd0.yml
+- lib/lucky_sneaks/unidecoder_data/xd1.yml
+- lib/lucky_sneaks/unidecoder_data/xd2.yml
+- lib/lucky_sneaks/unidecoder_data/xd3.yml
+- lib/lucky_sneaks/unidecoder_data/xd4.yml
+- lib/lucky_sneaks/unidecoder_data/xd5.yml
+- lib/lucky_sneaks/unidecoder_data/xd6.yml
+- lib/lucky_sneaks/unidecoder_data/xd7.yml
+- lib/lucky_sneaks/unidecoder_data/xf9.yml
+- lib/lucky_sneaks/unidecoder_data/xfa.yml
+- lib/lucky_sneaks/unidecoder_data/xfb.yml
+- lib/lucky_sneaks/unidecoder_data/xfc.yml
+- lib/lucky_sneaks/unidecoder_data/xfd.yml
+- lib/lucky_sneaks/unidecoder_data/xfe.yml
+- lib/lucky_sneaks/unidecoder_data/xff.yml
+- lib/stringex.rb
+- stringex.gemspec
+has_rdoc: true
+homepage: http://github.com/rsl/stringex
+licenses: []
+
+post_install_message: 
+rdoc_options: 
+- --main
+- README.rdoc
+- --charset
+- utf-8
+- --line-numbers
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+requirements: []
+
+rubyforge_project: stringex
+rubygems_version: 1.3.2
+signing_key: 
+specification_version: 3
+summary: "Some [hopefully] useful extensions to Ruby\xE2\x80\x99s String class"
+test_files: 
+- test/acts_as_url_test.rb
+- test/string_extensions_test.rb
+- test/unidecoder_test.rb